November 2020 - mpiwg-tools - mailman3.cels.anl.gov

No call today
by Marc-André Hermanns 26 Nov '20

26 Nov '20

Dear all, due to Thanksgiving in the US, we will have no call today. Cheers, Marc-Andre -- Dr. rer. nat. Marc-André Hermanns IT Center Group: High Performance Computing Division: Computational Science and Engineering RWTH Aachen University Seffenter Weg 23 52074 Aachen Phone: +49 241 80-24381 hermanns(a)itc.rwth-aachen.de www.itc.rwth-aachen.de Social Media Kanäle des IT Centers: https://blog.rwth-aachen.de/itc/ https://www.facebook.com/itcenterrwth https://www.linkedin.com/company/itcenterrwth https://twitter.com/ITCenterRWTH https://www.youtube.com/channel/UCKKDJJukeRwO0LP-ac8x8rQ

1 0

Re: [mpiwg-tools] Intel MPI MQD support
by Durnov, Dmitry 24 Nov '20

24 Nov '20

Hi Joshua, I represent Intel MPI team here. libtvmpi.so should be part of all IMPI 2018 releases. (I checked up to IMPI 2018 U5) $ mpirun -V Intel(R) MPI Library for Linux* OS, Version 2018 Update 5 Build 20190404 (id: 18839) Copyright 2003-2019 Intel Corporation. $ objdump -x $I_MPI_ROOT/lib64/libtvmpi.so | grep mqs_ 000000000002c7b3 g F .text 0000000000000061 mqs_next_communicator 000000000002c357 g F .text 00000000000000e7 mqs_setup_process 000000000002c6d4 g F .text 000000000000006a mqs_setup_communicator_iterator 000000000002baa1 g F .text 00000000000008b6 mqs_image_has_queues 000000000002b9c8 g F .text 000000000000000b mqs_dll_taddr_width 000000000002c619 g F .text 000000000000007e mqs_dll_error_string 000000000002c697 g F .text 000000000000003d mqs_update_communicator_list 000000000002ba19 g F .text 0000000000000088 mqs_setup_image 000000000002c9a9 g F .text 000000000000003a mqs_destroy_process_info 000000000002c73e g F .text 0000000000000075 mqs_get_communicator 000000000002c43e g F .text 00000000000001db mqs_process_has_queues 000000000002c9e3 g F .text 0000000000000022 mqs_destroy_image_info 000000000002c814 g F .text 00000000000000b3 mqs_setup_operation_iterator 000000000002b9bb g F .text 000000000000000d mqs_version_string 000000000002c8c7 g F .text 00000000000000e2 mqs_next_operation 000000000002b9d3 g F .text 0000000000000046 mqs_setup_basic_callbacks 000000000002b9b0 g F .text 000000000000000b mqs_version_compatibility As far as I remember we were porting that part to IMPI 2019/2021 as well but you are right that it isn't included into public package for some reason. I'm going to double check the situation. Thanks for the heads up. BR, Dmitry From: mpiwg-tools <mpiwg-tools-bounces(a)lists.mpi-forum.org> On Behalf Of Josh Cottingham via mpiwg-tools Sent: Tuesday, November 24, 2020 7:36 PM To: mpiwg-tools(a)lists.mpi-forum.org Cc: Josh Cottingham <Josh.Cottingham(a)arm.com>; Kevin Mooney <Kevin.Mooney(a)arm.com> Subject: [mpiwg-tools] Intel MPI MQD support Hey everyone, Is anyone within this group a part of the Intel MPI team or know a contact who works in that team? I was wondering if Message Queue Debugging support has been officially removed from Intel MPI? They've removed references to Message Queue Debugging from their user guide and no longer ship the message queue library libtvmpi.so. We cannot seem to find the find the mqs_ API anywhere (checked with Intel MPI newer than 2018.2). All the best, Joshua Cottingham | Software Engineer SDHE - Development Solutions Josh.Cottingham(a)arm.com<mailto:[email protected]> Arm IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you. -------------------------------------------------------------------- Joint Stock Company Intel A/O Registered legal address: Krylatsky Hills Business Park, 17 Krylatskaya Str., Bldg 4, Moscow 121614, Russian Federation This e-mail and any attachments may contain confidential material for the sole use of the intended recipient(s). Any review or distribution by others is strictly prohibited. If you are not the intended recipient, please contact the sender and delete all copies.

1 0

Intel MPI MQD support
by Josh Cottingham 24 Nov '20

24 Nov '20

Hey everyone, Is anyone within this group a part of the Intel MPI team or know a contact who works in that team? I was wondering if Message Queue Debugging support has been officially removed from Intel MPI? They've removed references to Message Queue Debugging from their user guide and no longer ship the message queue library libtvmpi.so. We cannot seem to find the find the mqs_ API anywhere (checked with Intel MPI newer than 2018.2). All the best, Joshua Cottingham | Software Engineer SDHE - Development Solutions Josh.Cottingham(a)arm.com Arm IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.

1 0

Re: [mpiwg-tools] Meeting reminder: Nov 19, 2020 (today) (CANCELLED)
by Marc-André Hermanns 19 Nov '20

19 Nov '20

Hi all, sorry for the noise (and short notice). I sent the reminder before recalling that some people still have conflicts with other calls, today. Therefore, let's skip today. I can use the time to do my RC reading assignment. Cheers, Marc-Andre On 19.11.20 16:41, Elis, Bengisu wrote: > Hi Marc-André, > > I wanted to let you know that neither Martin nor I will be able to join today's WG call because of other conflicting calls. > > Thank you and > Best, > - > Bengisu > > On 19.11.20, 16:33, "mpiwg-tools on behalf of Marc-André Hermanns via mpiwg-tools" <mpiwg-tools-bounces(a)lists.mpi-forum.org on behalf of mpiwg-tools(a)lists.mpi-forum.org> wrote: > > Hi all, > > This is a reminder of our regularly scheduled call today: > > Nov 19 at 8 am Pacific / 11 am Eastern / 5pm CEST. > > Webex Info: > https://github.com/mpiwg-tools/mpi-standard/wiki > > *NOTE*: The WebEx information moved to a non-public wiki. You have to > be logged into Github *and* have to have access to the private Tools > WG repositories. If you need access contact me. > > Agenda: > ------- > - QMPI > > > Cheers, > Marc-Andre > -- > Dr. rer. nat. Marc-André Hermanns > > IT Center > Group: High Performance Computing > Division: Computational Science and Engineering > RWTH Aachen University > Seffenter Weg 23 > 52074 Aachen > Phone: +49 241 80-24381 > hermanns(a)itc.rwth-aachen.de > www.itc.rwth-aachen.de > > > > > > > > > > > -- Dr. rer. nat. Marc-André Hermanns IT Center Group: High Performance Computing Division: Computational Science and Engineering RWTH Aachen University Seffenter Weg 23 52074 Aachen Phone: +49 241 80-24381 hermanns(a)itc.rwth-aachen.de www.itc.rwth-aachen.de Social Media Kanäle des IT Centers: https://blog.rwth-aachen.de/itc/ https://www.facebook.com/itcenterrwth https://www.linkedin.com/company/itcenterrwth https://twitter.com/ITCenterRWTH https://www.youtube.com/channel/UCKKDJJukeRwO0LP-ac8x8rQ

1 0

Meeting reminder: Nov 19, 2020 (today)
by Marc-André Hermanns 19 Nov '20

19 Nov '20

Hi all, This is a reminder of our regularly scheduled call today: Nov 19 at 8 am Pacific / 11 am Eastern / 5pm CEST. Webex Info: https://github.com/mpiwg-tools/mpi-standard/wiki *NOTE*: The WebEx information moved to a non-public wiki. You have to be logged into Github *and* have to have access to the private Tools WG repositories. If you need access contact me. Agenda: ------- - QMPI Cheers, Marc-Andre -- Dr. rer. nat. Marc-André Hermanns IT Center Group: High Performance Computing Division: Computational Science and Engineering RWTH Aachen University Seffenter Weg 23 52074 Aachen Phone: +49 241 80-24381 hermanns(a)itc.rwth-aachen.de www.itc.rwth-aachen.de

1 0

No call today
by Marc-André Hermanns 12 Nov '20

12 Nov '20

Dear all, this is just a reminder that we will skip today's call due to overlap with some SC20 workshops. Cheers, Marc-Andre -- Dr. rer. nat. Marc-André Hermanns IT Center Group: High Performance Computing Division: Computational Science and Engineering RWTH Aachen University Seffenter Weg 23 52074 Aachen Phone: +49 241 80-24381 hermanns(a)itc.rwth-aachen.de www.itc.rwth-aachen.de Social Media Kanäle des IT Centers: https://blog.rwth-aachen.de/itc/ https://www.facebook.com/itcenterrwth https://www.linkedin.com/company/itcenterrwth https://twitter.com/ITCenterRWTH https://www.youtube.com/channel/UCKKDJJukeRwO0LP-ac8x8rQ

1 0

Re: [mpiwg-tools] PMPI and sessions init
by William Gropp 10 Nov '20

10 Nov '20

Yes. We do not assume POSIX or Unix, and we try to stay away from specifics that can impact scalability (e.g., does this *semantically* require every process or node in a 10000000 node system to independently load files or save a GB of address space). Some things do, and some do not. But great care is needed to avoid tying MPI to particular implementations. Bill William Gropp Director, NCSA Thomas M. Siebel Chair in Computer Science University of Illinois Urbana-Champaign > On Nov 10, 2020, at 9:03 AM, Wesley Bland via mpiwg-tools <mpiwg-tools(a)lists.mpi-forum.org> wrote: > > We’ve explicitly avoided requiring specific mechanisms like dlopen so we can support the widest variety of tools on the widest variety of systems. dlopen, for instance, won’t work with static libraries, right? > > Thanks, > Wes > >> On Nov 9, 2020, at 8:11 AM, Joachim Protze via mpiwg-tools <mpiwg-tools(a)lists.mpi-forum.org <mailto:[email protected]>> wrote: >> >> If dlopen causes libdl to load a library (i.e., the library is not yet >> loaded and can be found and loaded), libdl calls the library constructor >> as part of dlopen. >> >> Forcing the MPI library to load (dlopen) all tool libraries before the >> registration window closes, gives all dynamic tools the chance to call >> the registration function, before MPI will continue initialization. > > _______________________________________________ > mpiwg-tools mailing list > mpiwg-tools(a)lists.mpi-forum.org > https://lists.mpi-forum.org/mailman/listinfo/mpiwg-tools

1 0

Re: [mpiwg-tools] PMPI and sessions init
by Wesley Bland 10 Nov '20

10 Nov '20

We’ve explicitly avoided requiring specific mechanisms like dlopen so we can support the widest variety of tools on the widest variety of systems. dlopen, for instance, won’t work with static libraries, right? Thanks, Wes > On Nov 9, 2020, at 8:11 AM, Joachim Protze via mpiwg-tools <mpiwg-tools(a)lists.mpi-forum.org> wrote: > > If dlopen causes libdl to load a library (i.e., the library is not yet > loaded and can be found and loaded), libdl calls the library constructor > as part of dlopen. > > Forcing the MPI library to load (dlopen) all tool libraries before the > registration window closes, gives all dynamic tools the chance to call > the registration function, before MPI will continue initialization.

1 0

Re: [mpiwg-tools] PMPI and sessions init
by Wesley Bland 10 Nov '20

10 Nov '20

> @Wesley: Why exactly did we come up with the restriction of "no > registration after init"? And did we push off the callback solution to > a later time? There’s a number of reasons, but the main one was that it makes the tool stack unpredictable and prevents the tool from caching the next functions in the call stack. It also gives us a point at which we can "turn on” interception of the QMPI_Register_function call so that meta-tools can intercept registration and start creating their own call stacks. Thanks, Wes

1 0

Re: [mpiwg-tools] PMPI and sessions init
by Joachim Protze 09 Nov '20

09 Nov '20

Hi Marc-Andre, I think, the point I missed to transport: If dlopen causes libdl to load a library (i.e., the library is not yet loaded and can be found and loaded), libdl calls the library constructor as part of dlopen. Forcing the MPI library to load (dlopen) all tool libraries before the registration window closes, gives all dynamic tools the chance to call the registration function, before MPI will continue initialization. I hope this makes is clearer - Joachim -- Dipl.-Inf. Joachim Protze IT Center Group: High Performance Computing Division: Computational Science and Engineering RWTH Aachen University Seffenter Weg 23 D 52074 Aachen (Germany) Tel: +49 241 80- 24765 Fax: +49 241 80-624765 protze(a)itc.rwth-aachen.de www.itc.rwth-aachen.de

1 0